The central goal of this project is to understand what causes indolent prostate cancer to become life threatening and metastasize to different parts of the body. To understand this critical process, we explore genetic data from human prostate cancer studies and recreate prostate cancer in genetically engineered mice. Until recently, it has not been possible to generate simply detectable metastatic prostate cancer with this approach and complex breeding issues combined with the artifact of massive lethal primary disease have severely hampered progress. We have solved these problems by directly transferring genes of interest into mouse prostate, generating a model system that we term RapidCaP. Now, we are in position to dissect the relevance of new prostate cancer gene candidates in both the classic model of cancer initiation as well as in RapidCaP. We are using this new system to analyze metastasis on two levels. At the molecular level, we correlate changes in protein expression with histological features of metastasis. This analysis has revealed that, paradoxically, Pten/ p53-mutant prostate metastasis does not show activation of the downstream Akt oncogene, but activation of Myc instead. Thus, we expand the RapidCaP approach to study how these molecular changes affect the pro- pensity for metastatic spread of disease in an animal. To validate these findings in human, we collaborate with clinicians that perform the corresponding molecular histology analysis on human metastatic prostate cancer samples. At the same time, we interrogate metastasis biology by engineering novel genetic changes that may cause metastasis in the RapidCaP model, as well as by analysis of the spontaneously occurring alterations that are associated with metastasis. This project aims therefore to combine discovery of candidate metastasis genes with their functional validation in the RapidCaP model, which can afford us with fundamental insights into disease progression. The derived results and hypotheses will then be validated in human by clinical collaborators as part of their separate, independently funded project on their biobanked human tissues. As a case in point, we are validating the role of a candidate tumor suppressor, the Phlpp2 gene, which effectively blocks Akt signaling. We find that this gene is regulated by p53 as well as by the Myc oncogene, which appears to use it to shut down Akt in metastasis. Therefore, we will test the role of this p53- and Myc-regulated gene using the RapidCaP and classic conditional knockout approaches in combination with knockout cells, and the results will be validated by collaborators using their human tissue from a separately funded project. Thus, I strongly believe that our proposed work on developing this novel system for the study of endogenous prostate metastasis in mouse can greatly accelerate the research on how prostate cancers become lethal.